- higherinter observerreliability = research is highly operationalised, several observers watching and coding the samebehaviour and agreeing on attachmentclassifications. Because of this observers have a clear view of for example how a securely attached infant should behave due to Ainsworthscriteria
- highly controlledobservation so easily replicated with the same results obtained
- standarddiagnostictool used to measure the strength and type of relationships between mother and infant
- highgeneralisability as it's been replicated by many cultures
- Low ecological validity = experiment was conducted in a controlledobservation - infant may be unfamiliar with surroundings so more likely distressed
- categories not always applicable (nomothetic) (mainandcassidy 1998) suggested that infants donot all fit into the 3 categories introduced by ainsworth ormaynotatall. They found another type of attachment (disorganised)
- culturally biased = SS was conducted in the USA. criteria used to criticise infants are based on USvalues e.g Japanese children were seen more as insecureresistant (spent most of their time with mother) whereas german infants were insecureavoidant (germans value independence)
- observation was a snapshot of behaviour, not taking into consideration of other factors like child's relationship with other family members or behaviour in familiar environments